US20040163000A1 - Network processor power management - Google Patents
Network processor power management Download PDFInfo
- Publication number
- US20040163000A1 US20040163000A1 US10/367,067 US36706703A US2004163000A1 US 20040163000 A1 US20040163000 A1 US 20040163000A1 US 36706703 A US36706703 A US 36706703A US 2004163000 A1 US2004163000 A1 US 2004163000A1
- Authority
- US
- United States
- Prior art keywords
- utilization
- processing element
- power
- power state
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012545 processing Methods 0.000 claims abstract description 104
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 230000008859 change Effects 0.000 claims abstract description 15
- 230000004044 response Effects 0.000 claims abstract description 9
- 238000012544 monitoring process Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 36
- 238000011084 recovery Methods 0.000 claims description 11
- 238000009499 grossing Methods 0.000 claims description 4
- 238000013461 design Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 32
- 239000000872 buffer Substances 0.000 description 19
- 238000007726 management method Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 13
- 230000009471 action Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 8
- 230000001276 controlling effect Effects 0.000 description 5
- 230000000737 periodic effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000001413 cellular effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000002411 adverse Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 208000013641 Cerebrofacial arteriovenous metameric syndrome Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000000593 degrading effect Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/02—Details
- H04L12/10—Current supply arrangements
Abstract
Description
- The present invention relates to a system and method of controlling power consumption within a network processor while maintaining an acceptable level of performance.
- Several trends in the communications industry are contributing to make the efficient use of power a top issue in the design of network system components. As bridging, switching and routing systems increase in performance, their power requirements also increase. An increase in power adversely affects product features, such as cost and reliability. Costs are increased, for example, by the requirements of larger power supplies and fans. Reliability is compromised by the potential of hotter components. The power increase also adversely affects operating environment features by driving higher utility costs and higher installation and maintenance costs, such as building cooling, space needs and battery backup requirements. On the other hand, network system components which run cool provide many important benefits, including the ability to pack more ports into a smaller space and still stay within thermal operating limits, and the capability to stay online longer, perhaps with reduced capacity, in a battery back-up mode when main power fails.
- Previously wired solutions for network communication are moving to wireless equipment for ease of use and mobility. Wireless devices abound today and will only increase in the future via analog and digital handheld phones, personal digital assistants (PDAs), laptops, subnotebooks, electronic books, etc. Most of these devices presently communicate or will communicate using wireless technologies, such as cellular, digital/PCS, 802.11, Bluetooth™, etc. Internet access is being enabled on most of these devices today or in the near future. New, data intensive features like web browsing and instant messaging are being added just as fast as improvements in low power hardware integration will allow. Network system components, such as network processors (NP) now used in powerful routing equipment today, may be used in small, mobile devices in the future provided that the technology is properly designed for low power applications.
- The amount of power that a device uses often varies greatly between passive use (such as a cellular phone in standby mode or a laptop computer in sleep mode) and active use (such as placing a call on a cellular phone or running an application on a laptop computer). Power management features allow a device to conserve power using different operational modes, such as standby or sleep. In these types of power modes, most of a device can be powered off with the state saved, parts of the device can be powered off with allowance for a wake-up by another part of the device, or parts of the device can be run at lower power during periods of low usage. These power-conserving operational modes can be used to greatly increase the battery life of the device and the amount of time that the device can be in standby or can be used actively.
- Due to the increasingly networked nature of wireless battery-powered devices (such as higher bandwidth requirements for multimedia and the addition of Internet support and applications), there is more of a need than ever to add networking assists or network processing functionality into these devices while leaving the General Purpose Processor (GPP) free to run applications. This drives the need for power-efficient network processors (NPs) and the use of power-saving techniques in the design of these network processors.
- These same power-saving features can be used in wired devices for battery-backup modes when there is a loss of power. An example would be a small-office/home-office (SOHO) router which is used to provide voice lines to a residence using Voice over Packet or Voice over ATM (asynchronous transfer mode) technology. In the United States, such a device must provide eight hours of “talk” time and 24hours of standby from a battery in case of a power failure to ensure 911 emergency support.
- These features can also be used in wired devices just to save electricity, which lowers operational expenses and is environmentally friendly. The use of advanced power management techniques according to the present invention will also improve thermal characteristics for high density network processor applications in telco racks by reducing typical power requirements. Carrier companies are asking for higher port densities to handle rapidly increasing volumes of voice and data traffic. These same companies are very sensitive to increases in overhead expenses due to floor space requirements for network equipment racks. It is becoming increasingly important for network equipment manufacturers to pack more network traffic processing capability into smaller spaces. Hence, the network processors used in this network equipment need to become more power efficient.
- The present invention relates to a programmable state machine included in an NP core for the purpose of monitoring the utilization of different processing elements in the NP and controlling the power state of each element as a function of past and predicted utilization. The invention utilizes a PM state control algorithm, run by a hardware or software state machine for determining when a processing element is to be set to a lower power state based on a historical measurement of a device and/or system utilization. State control logic accounts for dependencies between system elements before adjusting power states. Each processing element employs a usage indication signal that feeds a counter which is used as the algorithm that measures utilization levels. The invention is applicable to a centralized power management (PM) control unit, as well as a distributed PM approach, where each processing element includes its own state machine.
- The invention relates to a system and method of controlling power consumption within a network processor while maintaining an acceptable level of performance. The method utilizes a programmable state machine that is added to an NP core. The state machine is capable of monitoring the utilization of different processing elements in the NP and controlling the power state of each element as a function of past and predicted future utilization. The system can further include the capability of determining when a processing element is to be set to a lower power state, for example, based upon the prior utilization of the processing element, or upon the interdependencies between processing elements. For elements that are clocked, the performance recovery time associated with each of the clocked processing elements is factored into the step of controlling the power state of the respective processing element. The system can also include a smoothing step to account for the processing element recovery time. Gain factors can be adjusted in response to actual performance and utilization of the network processor.
- The invention also relates to a computer-readable medium for managing the power consumption of a system of multiple processing elements. The medium determines when each element in the system should be run based on a historical pattern of system utilization. It factors element interdependencies into power consumption requirements of each element and records utilization levels of each processing element.
- FIG. 1 is a block diagram of an integrated power management unit;
- FIG. 2 is a flow diagram illustrating the generation of NP power management state information;
- FIG. 3 is a flow diagram illustrating a process for sampling the utilization of processing elements within the network processor;
- FIG. 4 is a block diagram illustrating a system configuration for setting the power mode of a component based on the frequency of accesses;
- FIG. 5 is a flow chart illustrating the computation of threshold frequency using prior access frequencies stored in a ring buffer;
- FIG. 6 is a flow chart illustrating the computation of the access frequency and comparison to a threshold frequency for determining when to enter a power-save mode;
- FIG. 7 is a flow chart showing the use of the PM state controller to activate and deactivate power control signals to packet processors; and
- FIG. 8 is a magnetic disc representing the power consumption management medium of the present invention.
- A network processor is embedded in a semiconductor chip within an NP complex. The complex includes off-chip storage for the NP including DRAM and SRAM storage devices and data storage. It also supports a plurality of physical layer devices, such as 10/100 Mbps and 1 Gbps Ethernet, 16 Mbps Token-Ring, 200 Mbps fiber channels, SONET (synchronous optical network) and the like.
- The functional blocks of such a network processor are shown and described in greater detail on pages 27-39 of a public document entitled “IBM PowerNP™ NP4GS3 Network Processor”, the relevant portions of which are incorporated herein, and made a part hereof. The NP consists of several different processing elements, each of which has its own characteristic power and performance curve. For instance, an NP typically includes one imbedded control point processor (CPP) shown in FIGS.1-2 as the Embedded 405 PowerPC. The CPP handles initialization, table updates and special packet processing tasks. Assuming this is a CMOS (complimentary metal-oxide semiconductor) device, its power may be controlled using several methods already being employed in the mobile computing industry, i.e. clock throttling, clock stop and input voltage reduction. All of these power control methods may be used by the PM state machine; however, the control algorithm must take into account the performance recovery time associated with each method and the impact that a reduction in control processor performance can have on all other processing elements of the NP.
- Another NP processing element is the general purpose packet processor. An NP will typically contain dozens or even hundreds of these processors as a means of boosting the performance of the NP by spreading the packets across the packet processors in a multiprocessing approach. Each of these processors has an input queue associated with it such that the utilization of the packet processor can be determined by looking at the arrival rate of packets into the queue. A packet dispatch unit in the NP distributes the packet workload evenly across all of the packet processors. The collective utilization of all of the packet processors, therefore, can be approximated by measuring the utilization of one or a few processors. Also, assuming that the packet processors are identical in function, i.e. homogeneous, they are treated as a single entity and their total power consumption is regulated by starting and stopping individual processors. A calculated utilization threshold can be used by the PM state machine algorithm to determine when additional packet processors must be started or when packet processors are not needed and may be stopped. For the NP design case where all packet processors are not homogeneous, then groups of packet processors of the same type may be treated as a single processing entity in the same way.
- Other continuously clocked, homogeneous elements within an NP may be controlled in the same fashion, such as packet memory control units, scheduling queues, FIFO devices and Ingress/Egress port interfaces.
- Some external interfaces to the NP, such as MAC (media access control) port interfaces with free running clocks, are typically shut down when not in use. The PM state machine can shut down clocks and disable drivers to unused ports. It can also monitor certain interface signals which, when activated by an external attachment, indicate that the port should be enabled and that the interface clock should be started.
- All clocked processing elements within the NP have a recovery time associated with them which specifies the time required to switch from a low power state to a higher power or full power state. The recovery time for each type of processing element must be factored in the algorithms used by the PM state machine in order to adjust the reaction time to changes in utilization (first derivative of the control function) in order to minimize negative performance impacts. A smoothing function must be used in the control algorithm to account for the processing element recovery time. The PM state machine detects when an NP processing element is accessed and uses this information to calculate a current access frequency. The current access frequency or utilization value is compared to a previously calculated and continuously updated threshold frequency. The threshold frequency is representative of the access pattern, e.g., uniform or sporadic, and is computed from equations that include adjustable gain factors. During operation of the processing element, the appropriate power-save mode is entered when the current access frequency falls below the threshold frequency. The PM state machine algorithm can also adapt dynamically to varying workload situations, thereby saving more energy without degrading performance. This is accomplished by adjusting the gain factors in response to tracking the actual performance/utilization of the NP, thereby changing the processing element threshold frequency. The same PM state machine algorithm also determines when to exit a power-save mode, i.e. re-start processing elements. Details of this algorithm are described in U.S. Pat. No. 5,954,820 entitled PORTABLE COMPUTER WITH ADAPTIVE DEMAND-DRIVEN POWER MANAGEMENT, the teachings of which are incorporated herein.
- The present invention describes a PM state machine which, when combined with the referenced PM state control algorithm, can (a) control the power state of any element within the network processor, (b) select from any number of intermediate power states for a single element, (c) control each processing element individually based on unique energy saving and recovery time characteristics for each element, and (d) account for the dependencies between processing elements when changing PM states.
- The integrated power management unit,
IPMU 100 shown in FIG. 1, is a programmable state machine which updates power management (PM) state information at regular time intervals for all processing elements within the network processor. The processing elements include the embedded control processors, general purpose packet processors, co-processors, CAMS and other function accelerators, memory control units, queue and FIFO elements, ingress and egress port interfaces, switch interfaces, and host control bus interfaces, such as PCI. Each processing element has associated with it, a usage indication signal, one or more power control signals, i.e. on, off, standby, and possibly control signals, which change its operational characteristics based on PM state changes to other processing elements. The list of processing elements which may be controlled by the IPMU is kept in a table along with threshold information for allowable PM states, current PM state and utilization statistics. This table is illustrated as the NP element PM Table 104 attached to thePM activity monitor 106. The PM activity monitor 106 collectsutilization information 114 from each processing element, feeds the information into a PM state change algorithm and compares the results to predetermined threshold values to determine the number of packet processors which need to be active, as well as the optimum speed/power state for those elements with programmable clock states, i.e. the control point processor. Adjustable parameters used to tune the algorithms run by the PM activity monitor 106 are stored in thefunction parameter store 108. - If a PM state change is warranted because of a change in NP utilization, then the NP element
PM state controller 110 is activated and the logic sets the appropriate power control signals 116 to the correct processing element(s). ThePM state controller 110 uses an NP element state control table 112 that contains information on which control signals to activate to one or more NP processing elements according to an action specified by thePM activity monitor 106. PM state changes to some processing elements can affect the operation of other processing elements. For example, theNP dispatcher unit 118 is affected by which packet processors are activated and packet processors are affected by the power state of tree search engines. Logic in thePM state controller 110 maintains the operational integrity of the NP for every PM action. - FIG. 2 shows the process of using utilization data from NP processing elements to generate NP power management state information. These processing elements are arranged in a typical NP generally in accordance with the description in the aforementioned document entitled “IBM PowerNP NP4GS3 Network Processor”. A direct example of the NP power management can be shown by the control of active packet processors within an NP based on per-unit packet activity.
- The PM activity monitor (106 in FIG. 1) periodically obtains at 202 the current utilization values of each processing element listed in the NP element PM table (104 in FIG. 1). The utilization value for a packet processor is determined by choosing a time window and counting the number of input queue accesses that occur within that time window. The PM activity monitor then determines the next PM action (204) by running the PM algorithms (208) using past and new workload statistics. The NP element PM state table is then updated (206) with current values. A new index pointer is then sent at 210 to the NP element power management state controller. Using the index pointer, a new entry is loaded at 212 into the NP element state control table. The updated table then sets updated status mode control signals (214) followed by updated power mode control signals (216). Then, the next processing element in the PM Table is selected (218) and its utilization values are obtained (202) as the entire cycle is repeated to update the NP power management state information.
- FIG. 3 shows a flow chart of the polling process of FIG. 2. The method for gathering utilization data could also be interrupt driven. The polling process is repeated or duplicated for each packet processor or other processing element that is being monitored in the NP. An interval counter is loaded and started at302. As long as a processing element is active, its status is constantly being checked. Each time it is checked, it causes one count to be subtracted from
decrement timer 304. If the timer does not reach zero, the status of the processor element is checked at 316. If the element is determined to be active at 318, the activity counter is incremented by one at 324. The incremented value is again routed to thedecrement timer 304 where the cycle is repeated. If, however, the processor element is not active, the value in thedecrement timer 304 is again reduced. When decrement timer reaches zero, a new activity counter value is stored in the next location in the ring buffer at 310. The activity counter is then reset at 312 and the interval timer is reloaded and restarted for the next processing element. - FIG. 4 shows a configuration for use in implementing the process of FIG. 3. The
counter 402,timer 404 andring buffer elements 406 may be implemented in hardware or software. Anaccess signal 408 is sent to thecounter 402 each time there is a single packet processor access from the input queue. The accesses are counted continuously and the utilization value in the counter is output as anaccess density signal 410 connected to a storage location in thering buffer 406. The end-of-window signal 412 from the timer determines when the utilization value is written to the next storage location in the ring buffer and when the counter is reset. Thering buffer 406 stores a running history of packet processor utilization values equal to the number of storage locations in the buffer. The minimum number of values is two; however, a larger number of utilization values insure greater statistical accuracy for the utilization threshold. In addition to the inputs from the timer and counter, thering buffer 406 hasinputs 414 to clear an entry and a ‘read’input 416. It also has anoutput 418 to be described more fully in connection with FIG. 5. The activity measurement algorithm illustrated by FIGS. 3 and 4 can be used to store utilization values for any processing element as long as that element has at least one status signal indicating when it is active and when it is not active. The threshold computation algorithm illustrated in FIG. 5 works with the activity values collected by the algorithm in FIG. 3 for an individual processing element. - The PM activity monitor (106 in FIG. 1) maintains a utilization threshold value for each processing element listed in the PM table, as well as a link to a ring buffer (406 in FIG. 4) for each power mode. The ring buffers contain past utilization values that are used to update a utilization threshold for a specific power mode. The activity monitor 106 gets
utilization values 202 and compares the current utilization value, ut, against the utilization threshold value, tf, to determine the next PM action (204 in FIG. 2). A PM action may or may not correspond to the processing element in the PM table that was just sampled. In the case of power management for packet processors, a utilization threshold crossing condition indicates that other packet processors should be brought online or offline depending on a high or low crossing. A crossing below the utilization threshold indicates that one or more packet processors should be taken offline as soon they are idle. A crossing above the utilization threshold indicates that one or more packet processors should be brought online. The definition of which packet processors are started or stopped based on the utilization threshold crossing for a specific power mode is contained in the logic of the NP element PM state controller (110 in FIG. 1). - It is possible to detect periodic accesses to the packet processors from their utilization values. Enter and exit power saving modes can then be selected in anticipation of the beginning and end of a periodic access. An example of a periodic access is a long burst of network traffic entering the NP through a network port. The packets arrive and are processed at regular intervals corresponding to the maximum bandwidth of the port and the distribution of packets across the packet processors by the NP packet dispatcher unit. The ability to exit a power saving mode and start up additional packet processors just prior to the beginning of a periodic access improves the NP performance by reducing packet processing latency at the start of the traffic burst. The ability to enter a power-save mode by stopping one or more packet processors after a periodic access ends, increases the energy savings since the power-save mode is entered with a shortest possible delay.
- In one embodiment, component access patterns are assumed to fall into two categories: uniform access patterns and sporadic access patterns. A uniform access pattern is defined as one for which the set of access frequencies is well defined statistically. For example, the mean and standard deviation of access frequencies may be computed. The access frequencies can be considered well defined if the standard deviation is some fraction of the mean. Otherwise, the access pattern is considered sporadic, i.e., not well defined by a mean and standard deviation. In the uniform case, the access pattern is considered to have ended if the observed access frequencies drop below the mean minus some multiple of the standard deviation. This is equivalent to choosing a probability that observed access frequency belongs in the observed access pattern. It is also possible to use a fraction of a minimal observed access frequency as an estimate of the end of the access pattern. The underlying principle is the use of the recent access frequencies to characterize the access pattern, and then the determination of a threshold frequency from this pattern. It is then assumed to be highly probable that the access pattern has ceased when the access frequency crosses this threshold frequency.
- FIG. 5 is a flow chart showing the process for computing and adjusting the utilization threshold, tf, for a power-save mode. The current utilization threshold value, tf, is obtained at502. The most recent utilization values, ut, are obtained as the
output 418 of thering buffer 406 in FIG. 4. The number of values in the ring buffer is compared with the minimum threshold value at 504. If the number of values does not exceed the minimum, and if the current utilization value is found at 508 to be greater than half the sample time window, (tw), the store/modify branch of the process continues to store the current utilization value in the ring buffer at 510. If the utilization value does not exceed one half of the threshold value, (tf/2), then the last stored utilization value in the ring buffer is modified instep 520. Atstep 522, the algorithm returns to the activity monitor 106 in FIG. 1 to fetch another threshold value, tf, for the next processing element mode in the sequence. This algorithm is used to check and update all threshold values for all processing elements listed in the NP element PM table 104 in FIG. 1 - If the number of utilization values in the ring buffer is determined at504 to exceed the minimum threshold value, the current utilization value, ut, is compared to the utilization threshold, tf, at 506. If ut>tf, then a utilization peak has not been reached yet. There is no need to compute a new utilization threshold and the slope flag is set to zero at 512 to clear it. If the ring buffer contains a sufficient number of utilization values, and a new utilization peak has been found indicated by a change in slope from ut>tf to ut<=tf, then a new threshold frequency, tf, is computed at 516.
- The down-slope flag is set to 1 at518 a to indicate whether the algorithm is searching for an increase in packet processor activity or a decrease in activity. A new utilization threshold is calculated only on the first occurrence of a utilization value which is lower than the utilization threshold after the previously sampled utilization value was greater than the utilization threshold. This condition indicates that a new local maximum has been found for packet processor utilization. Immediately after a new utilization threshold is calculated, the flag is set (518) indicating that NP activity is on a downward slope from a new peak. The flag remains set at 1 as long as subsequent utilization values are less than or equal to the utilization threshold. The utilization threshold remains constant as long as the flag is set, indicating a downward slope in packet processor utilization. As soon as a sampled utilization value is compared and found to be greater than the utilization threshold, then the algorithm assumes that activity is now on an upward slope and begins searching for a new maximum. In this way, the algorithm is always searching for a new peak utilization threshold for the processing element.
- The utilization threshold is computed using a simplified calculation for the mean and standard deviation of the utilization values contained in the ring buffer. Only the maximum and minimum values contained in the ring buffer are used in the calculations. Two formulas are used to calculate two different utilization thresholds, t1 and t2. One formula tends to produce the greater value when the difference between the largest and the smallest utilization values in the ring buffer is small, indicating that the packet processor activity is uniform. The other formula tends to calculate the greater value when there is a large difference between the maximum and minimum utilization values, indicating sporadic packet processor activity. The greater of the two calculated values becomes the new utilization threshold, tf. Two gain constants are used in the formulas to control the sensitivity of the algorithm, g1 and g2. An increase in the value of either gain constant results in a lower calculated utilization threshold for the corresponding formula. In this way, the threshold algorithm can be adjusted to tune the power management behavior.
- t1=(maxut+minut)/2−g1*(maxut−minut)
- t2=minut/g2
- tf=max(t1, t2)
- The monitor, in effect, takes active/idle statistics that it has collected over time for each NP element and runs those through utilization algorithms to determine the best PM state for each element which will optimize power and performance characteristics for the whole NP system. The utilization threshold will tend to be a low value when packet processing activity is sporadic which will tend to put the NP into a high performance mode of operation. The utilization threshold will tend to either increase or remain constant when packet processing activity is uniform. The performance/power mode of the NP will tend to level out at its most optimum setting while packet processing activity is uniform. The algorithm run by the monitor can be tuned through programmable gain constants held in the function parameter store. In this way, the NP system can be programmed for a range of power management operation from maximum performance to maximum power savings. The utilization threshold algorithm run by the monitor maintains different threshold values for every processing element and for every power mode.
- FIG. 6 shows the logic flow used by the PM activity monitor602 to find the lowest possible power mode for a single processing element. The process is capable of working with one or more power saving modes for a processing element and it assumes that there is a separate threshold value calculated for each mode. At
step 604, the threshold frequencies for all power modes, last mode (lastm), are set to 1. Then the process waits for the shortest window timer to time out atstep 606. Each power save mode is presumed to have a unique window time. Starting with the lowest power mode (highest power savings),maxm 608, the logic obtains (610) and compares (612) the current activity, ut, to the activity threshold, tf(m), for that specific power-save mode. If the current activity is equal to or more than the threshold value, then the current activity value is stored, a new threshold value is calculated instep 614 for the power mode and the next best power mode is checked (616). If there are no more power modes available for the processing element, then the process goes back alongpath 617 to the beginning to check for the lowest power mode again. If there are additional power modes, the next lower mode is selected at 618 and the current activity is compared to the threshold value. - If the current activity is less than the threshold value, (ut<tf(m)), then an action code is sent along
path 613 to the NP element PM state controller to program the new and most energy-saving power state atstep 620. The NP element PM state controller receives PM actions from the monitor which instruct it to program new PM states for individual NP processing elements, e.g. packet processors. Different types of NP elements may have different PM state transition requirements besides On and Off, i.e. On > Standby > Off or OnH > OnL > Off. These requirements are specified in the NP element PM table along with dynamic utilization thresholds for each PM state change and the current power mode of each processing element. If the current mode is maxmode,step 622 indicates that the process is done (626). If there are still power-save modes available for testing, the last mode is updated to mode +1 to show that all power-save modes less than or equal to the mode that was entered are no longer under consideration. Step 606 is then entered again. - FIG. 7 shows the process used by the PM state controller to activate or deactivate power control signals to the packet processors. Each action index,
IDX 704, received from the PM monitor corresponds to an entry in the NP element state control table 702. Each table entry identifies which power mode signals,PMx 706, and status mode, SMx signals 708, are activated for a specific PM action. At 716, the PM state controller loads the power mode and status mode values obtained from the table 702. Power mode signals set one or more packet processors to a new power state and status mode signals change the operating characteristics of other, working processing elements at 724 affected by the new power mode, such as thedispatch unit 718. The PM control process includes await loop 722 to check for anidle condition 720 of the processing element before setting the packet processor power state. The PM state controller signals theNP dispatch unit 718 using status mode signals to indicate which packet processors are online or offline. The dispatch unit is updated before the packet processor is programmed so that no new packets are dispatched to the packet processor while the PM state controller is setting the new power mode. - FIG. 8 shows a computer-readable medium in the form of a
floppy disc 800 for containing the software implementation of the program to carry out the various steps of managing the power consumption of a network processor according to the present invention. - While the invention has been described in combination with specific embodiments thereof, there are many alternatives, modifications, and variations that are likewise deemed to be within the scope thereof. Accordingly, the invention is intended to embrace all such alternatives, modifications and variations as fall within the spirit and scope of the appended claims.
Claims (26)
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/367,067 US7337334B2 (en) | 2003-02-14 | 2003-02-14 | Network processor power management |
KR1020030100387A KR100544402B1 (en) | 2003-02-14 | 2003-12-30 | Network Processor Power Management |
JP2004032733A JP2004246894A (en) | 2003-02-14 | 2004-02-09 | Power saving method and device for network processor |
TW093103026A TWI269963B (en) | 2003-02-14 | 2004-02-10 | Method, system and computer readable media of controlling the power state of multiple processing elements in a network processor |
US11/970,613 US7472293B2 (en) | 2003-02-14 | 2008-01-08 | Network processor power management |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/367,067 US7337334B2 (en) | 2003-02-14 | 2003-02-14 | Network processor power management |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US11/970,613 Continuation US7472293B2 (en) | 2003-02-14 | 2008-01-08 | Network processor power management |
Publications (2)
Publication Number | Publication Date |
---|---|
US20040163000A1 true US20040163000A1 (en) | 2004-08-19 |
US7337334B2 US7337334B2 (en) | 2008-02-26 |
Family
ID=32849888
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/367,067 Expired - Fee Related US7337334B2 (en) | 2003-02-14 | 2003-02-14 | Network processor power management |
US11/970,613 Expired - Fee Related US7472293B2 (en) | 2003-02-14 | 2008-01-08 | Network processor power management |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US11/970,613 Expired - Fee Related US7472293B2 (en) | 2003-02-14 | 2008-01-08 | Network processor power management |
Country Status (4)
Country | Link |
---|---|
US (2) | US7337334B2 (en) |
JP (1) | JP2004246894A (en) |
KR (1) | KR100544402B1 (en) |
TW (1) | TWI269963B (en) |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050138442A1 (en) * | 2003-12-22 | 2005-06-23 | International Business Machines Corporation | Method and system for energy management in a simultaneous multi-threaded (SMT) processing system including per-thread device usage monitoring |
US20060050729A1 (en) * | 2004-09-08 | 2006-03-09 | Paul Chen | Wireless state machine and multiplexing method for concurrent ad-hoc and infrastructure mode service in wireless networking |
US20060064679A1 (en) * | 2004-08-19 | 2006-03-23 | Matsushita Electric Industrial Co., Ltd. | Processing apparatus |
US20060149975A1 (en) * | 2004-12-30 | 2006-07-06 | Intel Corporation | Operating point management in multi-core architectures |
US20060218424A1 (en) * | 2005-03-23 | 2006-09-28 | Miron Abramovici | Integrated circuit with autonomous power management |
US20080188230A1 (en) * | 2007-02-07 | 2008-08-07 | Samsung Electronics Co., Ltd. | Method and apparatus for allocating radio resource in wireless sensor network |
CN100442204C (en) * | 2006-07-19 | 2008-12-10 | 北京天碁科技有限公司 | System-on-chip chip and its power consumption control method |
US7472306B2 (en) * | 2004-05-18 | 2008-12-30 | Intel Corporation | Processor timing apparatus, systems, and methods |
US20090006877A1 (en) * | 2007-06-28 | 2009-01-01 | Seagate Technology Llc | Power management in a storage array |
US20090006871A1 (en) * | 2007-06-28 | 2009-01-01 | Yen-Cheng Liu | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
US20090198979A1 (en) * | 2008-02-05 | 2009-08-06 | Dell Products L.P. | Processor performance state optimization |
US20090228724A1 (en) * | 2008-03-10 | 2009-09-10 | International Business Machines | Power Saving Method for Computer System |
US20090274070A1 (en) * | 2008-05-02 | 2009-11-05 | Shankar Mukherjee | Power management of networked devices |
US20100106436A1 (en) * | 2008-10-20 | 2010-04-29 | International Business Machines Corporation | Power consumption calculation facility |
US20110173360A1 (en) * | 2010-01-11 | 2011-07-14 | Qualcomm Incorporated | System and method of monitoring a central processing unit in real time |
EP2357757A1 (en) * | 2010-02-16 | 2011-08-17 | Comcast Cable Communications, LLC | System and method for capacity planning on a high speed data network |
WO2012003847A1 (en) | 2010-07-08 | 2012-01-12 | Nec Europe Ltd. | Method of supporting power control in a communication network |
US20120250571A1 (en) * | 2011-04-01 | 2012-10-04 | Daniel Matthew Andrews | Energy-efficient network device with coordinated scheduling and processor rate control |
US20130091365A1 (en) * | 2011-09-30 | 2013-04-11 | Poh Thiam Teoh | Robust governing of power management infrastructure in a bridge |
TWI455531B (en) * | 2011-09-02 | 2014-10-01 | Mediatek Inc | Network processor |
US8934993B2 (en) | 2010-10-27 | 2015-01-13 | Wistron Corporation | Life management circuit, an electronic system and a machine-implemented method for managing usage rates of multiple electronic components |
US9135126B2 (en) | 2013-02-07 | 2015-09-15 | International Business Machines Corporation | Multi-core re-initialization failure control system |
US20160195913A1 (en) * | 2006-12-29 | 2016-07-07 | Yen-Cheng Liu | Optimizing power usage by factoring processor architectural events to pmu |
US20160306416A1 (en) * | 2015-04-16 | 2016-10-20 | Intel Corporation | Apparatus and Method for Adjusting Processor Power Usage Based On Network Load |
US10437319B2 (en) * | 2011-03-11 | 2019-10-08 | Intel Corporation | Dynamic core selection for heterogeneous multi-core systems |
CN112947737A (en) * | 2021-02-20 | 2021-06-11 | 山东云海国创云计算装备产业创新中心有限公司 | Chip power consumption adjusting method and device, electronic equipment and storage medium |
US11054884B2 (en) * | 2016-12-12 | 2021-07-06 | Intel Corporation | Using network interface controller (NIC) queue depth for power state management |
US11429178B2 (en) * | 2019-02-19 | 2022-08-30 | Samsung Electronics Co., Ltd. | Electronic device and method for determining operating frequency of processor |
US11474588B2 (en) * | 2020-03-26 | 2022-10-18 | Research & Business Foundation Sungkyunkwan University | Method and apparatus for controlling power efficiency of processor based on polling I/O |
Families Citing this family (42)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060080461A1 (en) * | 2004-06-02 | 2006-04-13 | Wilcox Jeffrey R | Packet exchange for controlling system power modes |
JP4465598B2 (en) * | 2004-07-05 | 2010-05-19 | ソニー株式会社 | Integrated circuit, processing control method therefor, and program |
JP4846226B2 (en) * | 2004-10-26 | 2011-12-28 | 株式会社日立ソリューションズ | Information processing apparatus, information processing method, and program |
US20070124607A1 (en) * | 2005-11-30 | 2007-05-31 | Samsung Electronics Co., Ltd. | System and method for semi-automatic power control in component architecture systems |
US7516342B2 (en) * | 2005-12-30 | 2009-04-07 | Intel Corporation | Method, apparatus and system to dynamically choose an optimum power state |
US8929870B2 (en) | 2006-02-27 | 2015-01-06 | Qualcomm Incorporated | Methods, apparatus, and system for venue-cast |
US7415623B2 (en) * | 2006-07-31 | 2008-08-19 | Motorola, Inc. | System for managing the power source life between multiple individually powered devices in a wired system and method of using same |
US7949887B2 (en) | 2006-11-01 | 2011-05-24 | Intel Corporation | Independent power control of processing cores |
JP5182792B2 (en) * | 2007-10-07 | 2013-04-17 | アルパイン株式会社 | Multi-core processor control method and apparatus |
EP2232721B1 (en) * | 2007-12-14 | 2014-08-13 | Telefonaktiebolaget L M Ericsson (publ) | Power control optimization in a communication network |
US8650413B2 (en) * | 2009-04-15 | 2014-02-11 | International Business Machines Corporation | On-chip power proxy based architecture |
US8271809B2 (en) * | 2009-04-15 | 2012-09-18 | International Business Machines Corporation | On-chip power proxy based architecture |
US8650425B2 (en) | 2009-05-06 | 2014-02-11 | Via Technologies, Inc. | Computer system for processing data in non-operational state and processing method thereof |
JP5293396B2 (en) * | 2009-05-13 | 2013-09-18 | 日本電気株式会社 | Information processing apparatus and control method of information processing apparatus |
US9298583B2 (en) * | 2009-06-04 | 2016-03-29 | International Business Machines Corporation | Network traffic based power consumption estimation of information technology systems |
US8473818B2 (en) * | 2009-10-12 | 2013-06-25 | Empire Technology Development Llc | Reliable communications in on-chip networks |
GB0919253D0 (en) | 2009-11-03 | 2009-12-16 | Cullimore Ian | Atto 1 |
US20110145612A1 (en) * | 2009-12-16 | 2011-06-16 | International Business Machines Corporation | Method and System to Determine and Optimize Energy Consumption of Computer Systems |
TWI514127B (en) * | 2010-04-16 | 2015-12-21 | Via Tech Inc | A computer system with an e-reader mode and e-book processing method thereof |
CN103080899B (en) * | 2010-07-13 | 2016-07-06 | 超威半导体公司 | In graphic process unit, SIMD unit dynamically enables and disables |
US8693679B1 (en) * | 2010-07-22 | 2014-04-08 | Adtran, Inc. | Communications system and associated method for reducing power consumption of a telecommunications device |
US8564922B1 (en) | 2010-07-22 | 2013-10-22 | Adtran, Inc. | System and method for protecting telecommunications device from power faults |
US8875276B2 (en) * | 2011-09-02 | 2014-10-28 | Iota Computing, Inc. | Ultra-low power single-chip firewall security device, system and method |
TWI450084B (en) * | 2010-12-22 | 2014-08-21 | Via Tech Inc | Decentralized power management distributed among multiple processor cores |
US8533512B2 (en) * | 2011-02-10 | 2013-09-10 | International Business Machines Corporation | Dynamic power and performance calibration of data processing systems |
US8904216B2 (en) | 2011-09-02 | 2014-12-02 | Iota Computing, Inc. | Massively multicore processor and operating system to manage strands in hardware |
WO2013101193A1 (en) * | 2011-12-30 | 2013-07-04 | Intel Corporation | Method and device for managing hardware errors in a multi-core environment |
JP2013149221A (en) * | 2012-01-23 | 2013-08-01 | Canon Inc | Control device for processor and method for controlling the same |
US8984313B2 (en) | 2012-08-31 | 2015-03-17 | Intel Corporation | Configuring power management functionality in a processor including a plurality of cores by utilizing a register to store a power domain indicator |
JP5462920B2 (en) * | 2012-10-19 | 2014-04-02 | アラクサラネットワークス株式会社 | Packet transfer device |
TWI519922B (en) | 2013-06-07 | 2016-02-01 | 智邦科技股份有限公司 | Power saving device and power saving method thereof |
US9766686B2 (en) | 2014-01-14 | 2017-09-19 | Cisco Technology, Inc. | Dynamically adjusting data processing speed based on buffer utilization |
US9565032B2 (en) | 2014-12-03 | 2017-02-07 | Qualcomm Incorporated | Monitoring and controlling bus ring performance |
US9588574B2 (en) | 2015-01-28 | 2017-03-07 | Qualcomm Incorporated | Power saving mode fallback during concurrency scenarios |
US9374786B1 (en) | 2015-02-17 | 2016-06-21 | Qualcomm Incorporated | System and methods for improving opportunistic envelope tracking in a multi-subscriber identity module (SIM) wireless communication device |
JP6380261B2 (en) * | 2015-06-29 | 2018-08-29 | 京セラドキュメントソリューションズ株式会社 | Electronic equipment and power supply control program |
US9886081B2 (en) | 2015-09-16 | 2018-02-06 | Qualcomm Incorporated | Managing power-down modes |
US10740167B2 (en) * | 2016-12-07 | 2020-08-11 | Electronics And Telecommunications Research Institute | Multi-core processor and cache management method thereof |
CN107733979B (en) * | 2017-09-06 | 2021-05-28 | 中国科学院计算机网络信息中心 | Data pushing method, server and system |
CN107707331A (en) * | 2017-09-13 | 2018-02-16 | 深圳市鑫宇鹏电子科技有限公司 | WiFi chip and safety chip information interacting method |
KR102552666B1 (en) | 2018-11-12 | 2023-07-10 | 삼성전자주식회사 | Electronic device and operation method thereof |
US11126245B2 (en) * | 2019-06-21 | 2021-09-21 | Intel Corporation | Device, system and method to determine a power mode of a system-on-chip |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4312035A (en) * | 1979-06-18 | 1982-01-19 | Greene Richard E | Apparatus for controlling electrical power in a data processing system |
US4679191A (en) * | 1983-05-04 | 1987-07-07 | Cxc Corporation | Variable bandwidth switching system |
US5251320A (en) * | 1990-05-25 | 1993-10-05 | International Business Machines Corporation | Power controller for permitting multiple processors to power up shared input/output devices and inhibit power down until all processors have ceased service with the I/O devices |
US5469542A (en) * | 1991-07-22 | 1995-11-21 | International Business Machines Corporation | Serial diagnostic interface bus for multiprocessor systems |
US5675814A (en) * | 1996-01-26 | 1997-10-07 | Dell Usa, L.P. | Apparatus and method for managing power consumption by I/O ports in a computer system |
US5691948A (en) * | 1995-06-08 | 1997-11-25 | Fujitsu Limited | Memory apparatus |
US5737615A (en) * | 1995-04-12 | 1998-04-07 | Intel Corporation | Microprocessor power control in a multiprocessor computer system |
US5954820A (en) * | 1997-09-26 | 1999-09-21 | International Business Machines Corporation | Portable computer with adaptive demand-driven power management |
US5983357A (en) * | 1995-07-28 | 1999-11-09 | Compaq Computer Corporation | Computer power management |
US6125450A (en) * | 1996-12-19 | 2000-09-26 | Intel Corporation | Stop clock throttling in a computer processor through disabling bus masters |
US6128746A (en) * | 1997-08-26 | 2000-10-03 | International Business Machines Corporation | Continuously powered mainstore for large memory subsystems |
US6161187A (en) * | 1996-11-01 | 2000-12-12 | Compaq Computer Corporation | Skipping clock interrupts during system inactivity to reduce power consumption |
US20020049892A1 (en) * | 2000-03-08 | 2002-04-25 | Ashley Saulsbury | Computer processing architecture having a scalable number of processing paths and pipelines |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6785826B1 (en) | 1996-07-17 | 2004-08-31 | International Business Machines Corporation | Self power audit and control circuitry for microprocessor functional units |
US6657634B1 (en) * | 1999-02-25 | 2003-12-02 | Ati International Srl | Dynamic graphics and/or video memory power reducing circuit and method |
US20020133728A1 (en) * | 2000-11-14 | 2002-09-19 | Sanjay Agarwal | Network traffic based adaptive power management system for computer networks |
US6785827B2 (en) * | 2000-11-29 | 2004-08-31 | Dell Products L.P. | System for determining servers power supply requirement by sampling power usage values thereof at a rate based upon the criticality of its availability |
US7058824B2 (en) * | 2001-06-15 | 2006-06-06 | Microsoft Corporation | Method and system for using idle threads to adaptively throttle a computer |
US6986069B2 (en) * | 2002-07-01 | 2006-01-10 | Newisys, Inc. | Methods and apparatus for static and dynamic power management of computer systems |
US6885974B2 (en) * | 2003-01-31 | 2005-04-26 | Microsoft Corporation | Dynamic power control apparatus, systems and methods |
-
2003
- 2003-02-14 US US10/367,067 patent/US7337334B2/en not_active Expired - Fee Related
- 2003-12-30 KR KR1020030100387A patent/KR100544402B1/en not_active IP Right Cessation
-
2004
- 2004-02-09 JP JP2004032733A patent/JP2004246894A/en active Pending
- 2004-02-10 TW TW093103026A patent/TWI269963B/en not_active IP Right Cessation
-
2008
- 2008-01-08 US US11/970,613 patent/US7472293B2/en not_active Expired - Fee Related
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4312035A (en) * | 1979-06-18 | 1982-01-19 | Greene Richard E | Apparatus for controlling electrical power in a data processing system |
US4679191A (en) * | 1983-05-04 | 1987-07-07 | Cxc Corporation | Variable bandwidth switching system |
US5251320A (en) * | 1990-05-25 | 1993-10-05 | International Business Machines Corporation | Power controller for permitting multiple processors to power up shared input/output devices and inhibit power down until all processors have ceased service with the I/O devices |
US5469542A (en) * | 1991-07-22 | 1995-11-21 | International Business Machines Corporation | Serial diagnostic interface bus for multiprocessor systems |
US6202097B1 (en) * | 1991-07-22 | 2001-03-13 | International Business Machines Corporation | Methods for performing diagnostic functions in a multiprocessor data processing system having a serial diagnostic bus |
US5737615A (en) * | 1995-04-12 | 1998-04-07 | Intel Corporation | Microprocessor power control in a multiprocessor computer system |
US5691948A (en) * | 1995-06-08 | 1997-11-25 | Fujitsu Limited | Memory apparatus |
US5983357A (en) * | 1995-07-28 | 1999-11-09 | Compaq Computer Corporation | Computer power management |
US5675814A (en) * | 1996-01-26 | 1997-10-07 | Dell Usa, L.P. | Apparatus and method for managing power consumption by I/O ports in a computer system |
US6161187A (en) * | 1996-11-01 | 2000-12-12 | Compaq Computer Corporation | Skipping clock interrupts during system inactivity to reduce power consumption |
US6125450A (en) * | 1996-12-19 | 2000-09-26 | Intel Corporation | Stop clock throttling in a computer processor through disabling bus masters |
US6128746A (en) * | 1997-08-26 | 2000-10-03 | International Business Machines Corporation | Continuously powered mainstore for large memory subsystems |
US5954820A (en) * | 1997-09-26 | 1999-09-21 | International Business Machines Corporation | Portable computer with adaptive demand-driven power management |
US20020049892A1 (en) * | 2000-03-08 | 2002-04-25 | Ashley Saulsbury | Computer processing architecture having a scalable number of processing paths and pipelines |
Cited By (74)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7197652B2 (en) * | 2003-12-22 | 2007-03-27 | International Business Machines Corporation | Method and system for energy management in a simultaneous multi-threaded (SMT) processing system including per-thread device usage monitoring |
US20050138442A1 (en) * | 2003-12-22 | 2005-06-23 | International Business Machines Corporation | Method and system for energy management in a simultaneous multi-threaded (SMT) processing system including per-thread device usage monitoring |
US7472306B2 (en) * | 2004-05-18 | 2008-12-30 | Intel Corporation | Processor timing apparatus, systems, and methods |
US7594131B2 (en) | 2004-08-19 | 2009-09-22 | Panasonic Corporation | Processing apparatus |
US20060064679A1 (en) * | 2004-08-19 | 2006-03-23 | Matsushita Electric Industrial Co., Ltd. | Processing apparatus |
US20080276044A1 (en) * | 2004-08-19 | 2008-11-06 | Matsushita Electric Industrial Co., Ltd. | Processing apparatus |
US7680087B2 (en) * | 2004-09-08 | 2010-03-16 | Canon U.S.A., Inc. | Wireless state machine and multiplexing method for concurrent ad-hoc and infrastructure mode service in wireless networking |
US20060050729A1 (en) * | 2004-09-08 | 2006-03-09 | Paul Chen | Wireless state machine and multiplexing method for concurrent ad-hoc and infrastructure mode service in wireless networking |
US11287871B2 (en) | 2004-12-30 | 2022-03-29 | Intel Corporation | Operating point management in multi-core architectures |
US8650424B2 (en) | 2004-12-30 | 2014-02-11 | Intel Corporation | Method and apparatus to control power consumption of a plurality of processor cores |
US20110099397A1 (en) * | 2004-12-30 | 2011-04-28 | Efraim Rotem | Operating point management in multi-core architectures |
US20060149975A1 (en) * | 2004-12-30 | 2006-07-06 | Intel Corporation | Operating point management in multi-core architectures |
US10013047B2 (en) | 2004-12-30 | 2018-07-03 | Intel Corporation | Operating point management in multi-core architectures |
US7502948B2 (en) * | 2004-12-30 | 2009-03-10 | Intel Corporation | Method, system, and apparatus for selecting a maximum operation point based on number of active cores and performance level of each of the active cores |
US20090172375A1 (en) * | 2004-12-30 | 2009-07-02 | Intel Corporation | Operating Point Management in Multi-Core Architectures |
US9619009B2 (en) | 2004-12-30 | 2017-04-11 | Intel Corporation | Operating point management in multi-core architectures |
US9785226B2 (en) | 2004-12-30 | 2017-10-10 | Intel Corporation | Operating point management in multi-core architectures |
US20060218424A1 (en) * | 2005-03-23 | 2006-09-28 | Miron Abramovici | Integrated circuit with autonomous power management |
CN100442204C (en) * | 2006-07-19 | 2008-12-10 | 北京天碁科技有限公司 | System-on-chip chip and its power consumption control method |
US20160195913A1 (en) * | 2006-12-29 | 2016-07-07 | Yen-Cheng Liu | Optimizing power usage by factoring processor architectural events to pmu |
US11144108B2 (en) | 2006-12-29 | 2021-10-12 | Intel Corporation | Optimizing power usage by factoring processor architectural events to PMU |
US20080188230A1 (en) * | 2007-02-07 | 2008-08-07 | Samsung Electronics Co., Ltd. | Method and apparatus for allocating radio resource in wireless sensor network |
US8200234B2 (en) | 2007-02-07 | 2012-06-12 | Samsung Electronics Co., Ltd. | Method and apparatus for allocating radio resource in wireless sensor network |
GB2463193A (en) * | 2007-06-28 | 2010-03-10 | Intel Corp | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
US7814351B2 (en) * | 2007-06-28 | 2010-10-12 | Seagate Technology Llc | Power management in a storage array |
US7971074B2 (en) * | 2007-06-28 | 2011-06-28 | Intel Corporation | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
US20090006871A1 (en) * | 2007-06-28 | 2009-01-01 | Yen-Cheng Liu | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
WO2009006014A1 (en) * | 2007-06-28 | 2009-01-08 | Intel Corporation | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
GB2463193B (en) * | 2007-06-28 | 2012-04-25 | Intel Corp | Method, system, and apparatus for a core activity detector to facilitate dynamic power management in a distributed system |
US20090006877A1 (en) * | 2007-06-28 | 2009-01-01 | Seagate Technology Llc | Power management in a storage array |
US7992015B2 (en) | 2008-02-05 | 2011-08-02 | Dell Products L.P. | Processor performance state optimization |
US20090198979A1 (en) * | 2008-02-05 | 2009-08-06 | Dell Products L.P. | Processor performance state optimization |
US9304569B2 (en) | 2008-02-05 | 2016-04-05 | Dell Products L.P. | Processor performance state optimization |
US9274584B2 (en) | 2008-02-05 | 2016-03-01 | Dell Products L.P. | Processor performance state optimization |
US20090228724A1 (en) * | 2008-03-10 | 2009-09-10 | International Business Machines | Power Saving Method for Computer System |
US8214673B2 (en) * | 2008-03-11 | 2012-07-03 | International Business Machines Corporation | Power saving method for computer system |
US20090274070A1 (en) * | 2008-05-02 | 2009-11-05 | Shankar Mukherjee | Power management of networked devices |
US9454209B2 (en) | 2008-05-02 | 2016-09-27 | Dhaani Systems | Power management of networked devices |
US11853143B2 (en) | 2008-05-02 | 2023-12-26 | Dhaani Systems | Power management of networked devices |
US8488500B2 (en) | 2008-05-02 | 2013-07-16 | Dhaani Systems | Power management of networked devices |
WO2009134894A3 (en) * | 2008-05-02 | 2010-02-18 | Shankar Mukherjee | Power management of networked devices |
US11061461B2 (en) | 2008-05-02 | 2021-07-13 | Dhaani Systems | Power management of networked devices |
US20100106436A1 (en) * | 2008-10-20 | 2010-04-29 | International Business Machines Corporation | Power consumption calculation facility |
US9086877B2 (en) | 2010-01-11 | 2015-07-21 | Qualcomm Incorporated | System and method of monitoring a central processing unit in real time |
WO2011085323A3 (en) * | 2010-01-11 | 2011-12-22 | Qualcomm Incorporated | System and method of monitoring a central processing unit in real time |
CN104615229A (en) * | 2010-01-11 | 2015-05-13 | 高通股份有限公司 | System and method of monitoring a central processing unit in real time |
US8352759B2 (en) | 2010-01-11 | 2013-01-08 | Qualcomm Incorporated | System and method of monitoring a central processing unit in real time |
US20110173360A1 (en) * | 2010-01-11 | 2011-07-14 | Qualcomm Incorporated | System and method of monitoring a central processing unit in real time |
KR101388837B1 (en) | 2010-01-11 | 2014-04-23 | 퀄컴 인코포레이티드 | System and method of monitoring a central processing unit in real time |
CN102713789A (en) * | 2010-01-11 | 2012-10-03 | 高通股份有限公司 | System and method of monitoring a central processing unit in real time |
US20110199914A1 (en) * | 2010-02-16 | 2011-08-18 | Comcast Cable Communications, Llc | System and Method for Capacity Planning on a High Speed data Network |
US8797891B2 (en) | 2010-02-16 | 2014-08-05 | Comcast Cable Communications, Llc | System and method for capacity planning on a high speed data network |
EP2357757A1 (en) * | 2010-02-16 | 2011-08-17 | Comcast Cable Communications, LLC | System and method for capacity planning on a high speed data network |
US10187250B2 (en) | 2010-02-16 | 2019-01-22 | Comcast Cable Communications, Llc | System and method for capacity planning on a high speed data network |
US9106580B2 (en) | 2010-07-08 | 2015-08-11 | Nec Europe Ltd. | Method of supporting power control in a communication network |
WO2012003847A1 (en) | 2010-07-08 | 2012-01-12 | Nec Europe Ltd. | Method of supporting power control in a communication network |
US8934993B2 (en) | 2010-10-27 | 2015-01-13 | Wistron Corporation | Life management circuit, an electronic system and a machine-implemented method for managing usage rates of multiple electronic components |
US11755099B2 (en) | 2011-03-11 | 2023-09-12 | Intel Corporation | Dynamic core selection for heterogeneous multi-core systems |
US10437319B2 (en) * | 2011-03-11 | 2019-10-08 | Intel Corporation | Dynamic core selection for heterogeneous multi-core systems |
US10534424B2 (en) * | 2011-03-11 | 2020-01-14 | Intel Corporation | Dynamic core selection for heterogeneous multi-core systems |
US9258248B2 (en) * | 2011-04-01 | 2016-02-09 | Alcatel Lucent | Energy-efficient network device with coordinated scheduling and processor rate control |
US20120250571A1 (en) * | 2011-04-01 | 2012-10-04 | Daniel Matthew Andrews | Energy-efficient network device with coordinated scheduling and processor rate control |
TWI455531B (en) * | 2011-09-02 | 2014-10-01 | Mediatek Inc | Network processor |
US20130091365A1 (en) * | 2011-09-30 | 2013-04-11 | Poh Thiam Teoh | Robust governing of power management infrastructure in a bridge |
US9176554B2 (en) * | 2011-09-30 | 2015-11-03 | Intel Corporation | Power management in a bridge based on a second derivative value derived from a mathematical derivative of plurality of values corresponding to the received packets |
US9164853B2 (en) | 2013-02-07 | 2015-10-20 | International Business Machines Corporation | Multi-core re-initialization failure control system |
US9135126B2 (en) | 2013-02-07 | 2015-09-15 | International Business Machines Corporation | Multi-core re-initialization failure control system |
US20160306416A1 (en) * | 2015-04-16 | 2016-10-20 | Intel Corporation | Apparatus and Method for Adjusting Processor Power Usage Based On Network Load |
CN107430425A (en) * | 2015-04-16 | 2017-12-01 | 英特尔公司 | For adjusting the apparatus and method of processor power utilization rate based on network load |
US11797076B2 (en) | 2016-12-12 | 2023-10-24 | Intel Corporation | Using network interface controller (NIC) queue depth for power state management |
US11054884B2 (en) * | 2016-12-12 | 2021-07-06 | Intel Corporation | Using network interface controller (NIC) queue depth for power state management |
US11429178B2 (en) * | 2019-02-19 | 2022-08-30 | Samsung Electronics Co., Ltd. | Electronic device and method for determining operating frequency of processor |
US11474588B2 (en) * | 2020-03-26 | 2022-10-18 | Research & Business Foundation Sungkyunkwan University | Method and apparatus for controlling power efficiency of processor based on polling I/O |
CN112947737A (en) * | 2021-02-20 | 2021-06-11 | 山东云海国创云计算装备产业创新中心有限公司 | Chip power consumption adjusting method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
TWI269963B (en) | 2007-01-01 |
JP2004246894A (en) | 2004-09-02 |
US7337334B2 (en) | 2008-02-26 |
KR100544402B1 (en) | 2006-01-23 |
TW200502745A (en) | 2005-01-16 |
US20080109665A1 (en) | 2008-05-08 |
US7472293B2 (en) | 2008-12-30 |
KR20040073956A (en) | 2004-08-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20040163000A1 (en) | Network processor power management | |
US10429921B2 (en) | Datacenter power management optimizations | |
US7197652B2 (en) | Method and system for energy management in a simultaneous multi-threaded (SMT) processing system including per-thread device usage monitoring | |
US8516284B2 (en) | Saving power by placing inactive computing devices in optimized configuration corresponding to a specific constraint | |
Liu et al. | Sleepscale: Runtime joint speed scaling and sleep states management for power efficient data centers | |
KR101624765B1 (en) | Energy-aware server management | |
US7814485B2 (en) | System and method for adaptive power management based on processor utilization and cache misses | |
US8286013B2 (en) | Portable communication device with multi-tiered power save operation | |
US8261112B2 (en) | Optimizing power consumption by tracking how program runtime performance metrics respond to changes in operating frequency | |
US7346787B2 (en) | System and method for adaptive power management | |
US20050044435A1 (en) | System and method for reducing power consumption by estimating engine load and reducing engine clock speed | |
US20110106935A1 (en) | Power management for idle system in clusters | |
US9274585B2 (en) | Combined dynamic and static power and performance optimization on data centers | |
WO2010036592A2 (en) | Conserving power in a computer system | |
JP2002202893A (en) | Method for controlling execution of multiplex task and processing circuit | |
WO2008107344A2 (en) | Power management in a power-constrained processing system | |
WO2013090637A1 (en) | Dynamically modifying a power/performance tradeoff based on processor utilization | |
EP4160379A1 (en) | Storage system and energy-saving control method | |
Cho et al. | A battery lifetime guarantee scheme for selective applications in smart mobile devices | |
EP2112573B1 (en) | Power supply control method and circuit in communication equipment | |
CN103345296A (en) | Dynamic voltage frequency adjustment trigger device and method | |
Escheikh et al. | Workload-dependent software aging impact on performance and energy consumption in server virtualized systems | |
Nithi et al. | Smart power management for mobile handsets | |
Kargahi et al. | Stochastic DVS-based dynamic power management for soft real-time systems | |
US20050022037A1 (en) | Method, system, and apparatus for an efficient power dissipation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:KUHLMANN, CHARLES E.;MILLARD, THOMAS A.;STROLE, NORMAN C.;REEL/FRAME:013779/0846 Effective date: 20030213 |
|
FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
FPAY | Fee payment |
Year of fee payment: 4 |
|
REMI | Maintenance fee reminder mailed | ||
LAPS | Lapse for failure to pay maintenance fees | ||
STCH | Information on status: patent discontinuation |
Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362 |
|
FP | Lapsed due to failure to pay maintenance fee |
Effective date: 20160226 |